============================================================================
CODEBOOK: Idealpointestimates1946-2025.csv
============================================================================

TITLE
    United Nations General Assembly Ideal Point Estimates, 1946-2025

DESCRIPTION
    This dataset contains dynamic ideal point estimates of state preferences
    derived from United Nations General Assembly (UNGA) roll-call voting
    data. The estimates are produced using a Bayesian item-response theory
    (IRT) model with a random walk prior on state ideal points, estimated
    via Markov chain Monte Carlo (MCMC) methods. The model estimates a
    single latent dimension that captures state positions on a continuum
    where higher values indicate positions more aligned with the United
    States and Western democracies and lower values indicate positions
    more aligned with the developing world consensus.

METHOD
    Bailey, Michael A., Anton Strezhnev, and Erik Voeten. "Estimating
    Dynamic State Preferences from United Nations Voting Data." Journal
    of Conflict Resolution 61, no. 2 (2017): 430-456.

COVERAGE
    Temporal:   1946-2025 (79 UNGA sessions)
    Spatial:    197 unique countries/entities
    Unit:       Country-year
    Observations: 11,610

FILE FORMAT
    Comma-separated values (CSV), with header row.

============================================================================
VARIABLE DESCRIPTIONS
============================================================================

ccode
    Correlates of War (COW) numeric country code.
    Type: integer
    Range: 2-990

iso3c
    ISO 3166-1 alpha-3 country code.
    Type: string (3 characters)

Countryname
    Country name.
    Type: string

year
    Calendar year corresponding to the UNGA session.
    Type: integer
    Range: 1946-2025

NVotesFP
    Number of UNGA roll-call votes in which the country participated in
    the given year. This count reflects the votes used in the estimation
    after applying the "future projections" (FP) bridging procedure, which
    links votes across sessions to maintain comparability over time.
    Type: integer
    Range: 1-164

IdealPointFP
    Posterior mean of the estimated ideal point for the country-year,
    derived from the MCMC estimation. Higher values correspond to
    positions more aligned with the United States and Western democracies;
    lower values correspond to positions more aligned with the consensus
    of developing countries. The "FP" suffix denotes that estimates are
    produced using the future projections bridging procedure.
    Type: numeric (continuous)
    Range: approximately -3.59 to 3.36


Q5%FP
    5th percentile of the posterior distribution of the ideal point
    estimate.
    Type: numeric (continuous)

Q10%FP
    10th percentile of the posterior distribution of the ideal point
    estimate.
    Type: numeric (continuous)

Q50%FP
    Median (50th percentile) of the posterior distribution of the ideal
    point estimate.
    Type: numeric (continuous)

Q90%FP
    90th percentile of the posterior distribution of the ideal point
    estimate.
    Type: numeric (continuous)

Q95%FP
    95th percentile of the posterior distribution of the ideal point
    estimate.
    Type: numeric (continuous)


    Note: The interval [Q5%FP, Q95%FP] provides a 90% credible interval
    for the ideal point estimate. The interval [Q10%FP, Q90%FP] provides
    an 80% credible interval. Users can use these quantiles to assess the
    uncertainty surrounding each estimate.



============================================================================
NOTES
============================================================================

1. The ideal points are estimated on a latent scale and are comparable
   across countries and over time within this dataset. The dynamic model
   with bridging observations ensures intertemporal comparability.

2. Country-year observations appear only for years in which a country
   was a UN member state and cast at least one roll-call vote.

3. Estimates for country-years with very few votes (low NVotesFP) will
   have wider credible intervals, reflecting greater uncertainty.

4. Country codes follow the Correlates of War system. Users merging with
   other datasets should verify mappings, as COW codes can differ from
   other coding schemes (e.g., Gleditsch-Ward).

============================================================================
CITATION
============================================================================

    When using this dataset, please cite:

    Bailey, Michael A., Anton Strezhnev, and Erik Voeten. "Estimating
    Dynamic State Preferences from United Nations Voting Data." Journal
    of Conflict Resolution 61, no. 2 (2017): 430-456.

    Voeten, Erik. "United Nations General Assembly Ideal Point Estimates,
    1946-2025." Available at: https://dataverse.harvard.edu/dataverse/Voeten

============================================================================
